Is There Room for a New Desktop Operating System
10 comments
An interesting question, but you already know the answer to it :).
Linux has a lot of users giving it time and money and even it struggles with driver and other support. There are other desktop projects out there that take Amiga and BeOS and have a decent OS overall, but nobody can use them as serious daily drivers for work (ok, most of us can't). I think there is a lot of room for improvement here, but it's a massive project and so far nothing can touch those big three with the exception of BSD, which is the obvious 4th place I think.
What I really want is a billion dollars to have a brand new Smalltalk or Lisp machine built with great hardware support, graphics, and something like Mathematica builtin to it. Rebol/Red would also be good here.
Linux has a lot of users giving it time and money and even it struggles with driver and other support. There are other desktop projects out there that take Amiga and BeOS and have a decent OS overall, but nobody can use them as serious daily drivers for work (ok, most of us can't). I think there is a lot of room for improvement here, but it's a massive project and so far nothing can touch those big three with the exception of BSD, which is the obvious 4th place I think.
What I really want is a billion dollars to have a brand new Smalltalk or Lisp machine built with great hardware support, graphics, and something like Mathematica builtin to it. Rebol/Red would also be good here.
Hi,
I totally agree with your post. The issue with lacking drivers are often lacking specs (mostly for GPUs) so its not always the fault of the developer/project. I'm contributing code to Red myself and actually working on a PoC Kernel/OS implemented in Red/System. Its architecture will be an exoKernel/paraKernel. You would be able to directly boot into your IDE or app just like in the good old C64/Amiga (Game mode) days. What I really want to have is running Red/OS on the Mill CPU.
I totally agree with your post. The issue with lacking drivers are often lacking specs (mostly for GPUs) so its not always the fault of the developer/project. I'm contributing code to Red myself and actually working on a PoC Kernel/OS implemented in Red/System. Its architecture will be an exoKernel/paraKernel. You would be able to directly boot into your IDE or app just like in the good old C64/Amiga (Game mode) days. What I really want to have is running Red/OS on the Mill CPU.
First of all, thank you for your work. Since I don't contribute myself, I'm a bit of a parasite :)
Second, I'm salivating at the prospect of booting into a Red terminal directly on hardware. Would I be able to use the View DSL for graphics, or just the text parts of the console?
Best of luck! Also, I'd love to hear more details.
Second, I'm salivating at the prospect of booting into a Red terminal directly on hardware. Would I be able to use the View DSL for graphics, or just the text parts of the console?
Best of luck! Also, I'd love to hear more details.
Just keep using it and provide feedback :)
Yes, that's the plan, having full view DSL features. The issues are GPU drivers, as you don't get any specs from the manufacturers, so it will either be a software renderer or using a VGA/SVGA backend. An exoKernel/paraKernel is only securely multiplexing the resources. Everything happens in user-space (memory allocations, scheduling, etc.), so you can theoretically run every app from every OS at the same time (if you provide a libOS being able to handle the API calls of the specific OS). You would be able to run only one single app or multiple ones. The OS is just another application so you can boot into an OS or an application itself, running on bare-metal.
Yes, that's the plan, having full view DSL features. The issues are GPU drivers, as you don't get any specs from the manufacturers, so it will either be a software renderer or using a VGA/SVGA backend. An exoKernel/paraKernel is only securely multiplexing the resources. Everything happens in user-space (memory allocations, scheduling, etc.), so you can theoretically run every app from every OS at the same time (if you provide a libOS being able to handle the API calls of the specific OS). You would be able to run only one single app or multiple ones. The OS is just another application so you can boot into an OS or an application itself, running on bare-metal.
That is cool. Thanks!
Companies want proven systems with a lot of available support/know-how. Employers want systems that don't need much training for new employees (at least on the OS level). Most private customers don't care about user-serviceability and upgradability and aren't interested in learning something new on the OS level - they buy cheap notebooks with Windows or some Chromebooks or even iPads.
Leaves the enthusiasts: the gamers, the self-employed and SMB in the IT and media environment, musicians etc. They would love upgradability, but they need the software and the ecosystems they (and in case of the self-employed and SMB their customers) are used to. That was limiting the use of Linux from the beginning. It limits the use of Macs for many as well.
So: Because of the complete ecosystem (knowledge, programs, drivers, proven reliability) missing, anything new would have a very hard time to get traction.
S
Leaves the enthusiasts: the gamers, the self-employed and SMB in the IT and media environment, musicians etc. They would love upgradability, but they need the software and the ecosystems they (and in case of the self-employed and SMB their customers) are used to. That was limiting the use of Linux from the beginning. It limits the use of Macs for many as well.
So: Because of the complete ecosystem (knowledge, programs, drivers, proven reliability) missing, anything new would have a very hard time to get traction.
S
In short, probably not. IT departments want open source and existing desktop users mostly want Windows. Given Microsoft's recent rumblings, I think the next major thing we're going to see on the desktop is the full virtualization of Windows running under Linux and officially supported by MS. This solves both of the most pressing needs (get to open source and don't disrupt existing apps).
It would take a decade to build a good desktop OS.
We are soon entering the golden age of a browser. Even Microsoft Office is going to be software-as-a-service. There is no need for a desktop, as there are not going to be any desktop applications in the future. Eventually we all boot to Chrome that executes Wasm and WebGL in near native speeds.
We are soon entering the golden age of a browser. Even Microsoft Office is going to be software-as-a-service. There is no need for a desktop, as there are not going to be any desktop applications in the future. Eventually we all boot to Chrome that executes Wasm and WebGL in near native speeds.
> or the Linux desktop
Which one? There's a bunch of them.
Which one? There's a bunch of them.
Against the grain I know, but I have the inclination that, if you’re asking, then there likely is?
Since the demise of BeOS in the early 2000s, there hasn't been any commercial competitors to Windows and macOS in the desktop computing realm other than perhaps ChromeOS. However, unlike the early 2000s when Windows' marketshare was in the 90+ percent range, today there is more diversity of personal computing platforms thanks to the popularity of smartphones, tablets, and Chromebooks. Given this new diversity of platforms in 2019, I am wondering if there is room for a new commercial desktop operating system to emerge, one that attracts users that are frustrated with Apple's direction of the Mac but who don't want to switch to Windows or the Linux desktop.